1. Field of the Invention
The present invention relates to a seek control method, seek control device and disk device for controlling seeking of an object to a target position, and more particularly to a seek control method, seek control device and disk device for controlling seeking by a two-degree of freedom control system.
2. Description of the Related Art
A control for moving an object to a target position, that is so called “seek control”, is widely known. Particularly in disk devices, such as magnetic disk devices and optical disk devices, accurately positioning the head to a target track by seek control is extremely important to improve recording density.
In seek control for a disk device, a control method called two-degree of freedom control, or feed forward control, is used. In this control method, a target position trajectory is supplied to the control system according to the time elapsed from the seek start time, or both target position trajectory and current trajectory are supplied.
FIG. 25 is a block diagram depicting the control system of a conventional two-degree of freedom control. As FIG. 25 shows, when a plant 108, which is a control object, is moved to the target position “r” by seeking, the position trajectory, or velocity trajectory or acceleration trajectory uf is created in the trajectory creation block 100 from the target position “r”, and is supplied to the plant 108.
In the feed back loop system, the computing block 102 computes position error between the current position “y” of the plant 108 and the target position “r”, and the controller 104 computes a control amount so as to decrease the position error from the position error, and the computing block 106 adds the control amount to the above mentioned seek trajectory uf, and drives the plant 108 using this added value “u”.
In this way, the seek trajectory uf is created from the target position “r” to perform seek control. The characteristic Pm of the trajectory creation block 100 is ideally the same characteristic P of the plant 108. In other words, if the characteristic of the plant 108 is P, the characteristic of the trajectory creation block 100 is ideally 1/Pm=1/P.
To create a seek trajectory, a method of computing the current trajectory (acceleration trajectory) and position trajectory using a polynomial of the elapsed time from the seek start time has been proposed (e.g. Japanese Patent Applications Laid-Open No. H03-233609 and No. H05-143165). Also a method of generating a trajectory to remove the influence of resonance frequency is proposed (e.g. Japanese Patent Applications Laid-Open No. 2001-210033, No. H8-007506 and No. H6-187044).
In these prior arts, the object is suppressing the resonance of the plant by improving the seek trajectory so as to increase the speed of seeking. A problem of seek control is a shift between the target trajectory and the actual response of the plant. This shift is mainly generated by the gain difference between the model (block 100) when the target trajectory is generated and the actual control target 108. Currently as the track density increases, even a 1% shift causes several hundred tracks of shift if the seek distance reaches several tens of thousand tracks.
The feed back control loop of the prior art operates such that this position shift is corrected, but during seek control, the servo frequency band is dropped compared with the time of following so that the trajectory control is not interfered with. Therefore if the shift is major, an overrun or underrun is generated, and re-seeking becomes necessary, and seek time when seek ends disperses.